Abstract: This research proposes, develops, and demonstrates a quantitative risk analytic method for integrating a set of modern deterrence considerations with respect to nuclear weapon arsenals and policies. These considerations include multiple prospective models of antagonist behaviors, multiple levels of conflict escalation, multiple weapon capabilities and effects, and nuanced policies for protagonists and antagonists. A mathematical basis for this approach is developed on the foundation of infinite-horizon, risk-sensitive Interactive Partially Observable Markov Decision Processes (IPOMDP). This foundation allows multiple decision agents to identify optimal policies when managing conflict scenarios in the face of the tradeoff between achieving political goals and avoiding the consequences of various forms of conflict. A set of deterrence-effectiveness metrics that center on the probability of specific opponent actions and conflict outcomes occurring are suggested, and a method for evaluating them is proposed. The resulting modeling and analysis framework captures complex behaviors and escalation dynamics, identifies approximately optimal policies in specific conflict scenarios, and is extensible to a large array of possible conflict scenarios. An example analysis, based on fictitious data, analyzes a bilateral, nuclear-armed, peer-state competition in a conflict escalation scenario. The example analysis evaluates various nuclear weapons arsenals and stated employment policies by examining the optimal conflict management solutions produced by the method and by comparing deterrence-effectiveness metrics. The products of this research can serve as a foundation for future work to expand the model’s capabilities and enhance its performance. Most importantly, it will provide valuable insights to policy and decision makers in government.
Speaker Bio: Jason C. Reinhardt is a national security systems analyst and Distinguished Member of Technical Staff at Sandia National Laboratories. His work focuses on probabilistic analysis methods, quantitative and non-quantitative approaches for risk analysis and management, as well as the modeling and analysis of strategic interaction in conflict escalation, asymmetric deterrence, and stability. Jason received his Ph.D. in Risk Analysis from Stanford University School of Engineering’s Department of Management Science and Engineering. He also holds a M.S. in Electrical Engineering from Stanford University, and a B.S. in Electrical Engineering from the Purdue School of Electrical Engineering at Indianapolis.